基于改进ISODATA的海战场空间群形成研究

Research on Space-Group Formation on Sea Battlefield Based on Improved ISODATA

海战场空间群是从实体层目标信息中提取兵力层态势信息的基础,针对现有聚类算法的不足和可获取目标信息的特点运用改进ISODATA算法探索空间群形成问题。首先选取特征矢量与聚类测度,然后从改善多密度目标分群和初始类心优选2个方面对ISODATA进行改进,最后将聚类结果进行辅助识别,从而完成空间群形成建模。经仿真分析,该模型采用的新算法较改进前在分群有效性、适用性和运算效率上都有明显提升,较好地满足了空间群形成的需要,并可对其他背景下的聚类分群提供借鉴。

The space-group on the sea battlefield is the foundation of that distilling the force level situation information from the object level target information. Aiming at the deficiency of existing cluster algorithm and the characteristic of obtainable target information,this article researches the problem of space-group formation using the improved ISODATA. Firsdy,character vector and cluster measure are selected,secondly ISODATA is improved from two aspects which include amending multi-density clustering and original cluster center preference,lastly the clustering result is identified by assistant information to accomplish the space-group formation modeling. Compared with existing method,the improved algorithm has evident advance on the clustering validity, applicability and operation efficiency by simulation analyzing. This model can satisfy the requirement of space-group formation,and use for reference for clus＇tering under other background.